
The AI revolution isn't just about chips and algorithms. It's reshaping the physical infrastructure of data centres, creating opportunities for companies that can solve two fundamental problems: keeping servers cool and keeping them powered. Two Indian precision engineering firms—MTAR Technologies and Aeroflex Industries—are positioning themselves at the heart of this transformation.
The numbers are eye-catching. MTAR's standalone revenue jumped 200% year-on-year to ₹406.8 crore in Q2 FY27, while net profit surged roughly 14x to ₹59.8 crore. Aeroflex reported 87% revenue growth to ₹207.4 crore, with net profit up 157% to ₹36.6 crore. But these headline figures mask complex stories about customer concentration, capacity expansion, and the race to qualify for next-generation data centre infrastructure. Investor Presentation +2
MTAR's growth story is anchored in solid oxide fuel cells—components that provide continuous on-site power. The company has a 15-plus year relationship with a major international customer (widely reported as Bloom Energy, though MTAR doesn't name them) and holds "a majority share" of their requirements. Investor Presentation +1
The Clean Energy segment dominates MTAR's order book at 66.7%, with a closing book of ₹4,075.5 crore at end-Q2 FY27. Analysts on recent calls have linked potential customer expansions to hot box volumes of 60,000–77,000 units annually, though management declines to confirm specific numbers due to non-disclosure agreements. Investor Presentation +3
What's driving this demand? While MTAR's filings don't explicitly connect fuel cells to AI data centres, the logic is straightforward. AI data centres run 24/7 and often face grid connection delays. Solid oxide fuel cells provide reliable on-site power, making them attractive for hyperscalers building massive computing facilities. Management dismissed reports of US data centre capex delays as "unwanted noise" and said order flow is moving in the right direction. Transcript
The company is investing aggressively to keep pace. A ₹500 crore capex budget is roughly split 70:30 between clean energy and other segments. Capacity expansion is happening in three phases: Phase 1 is commissioned, Phase 2 is now expected by end-November 2026 (slipping from earlier September-October guidance), and Phase 3—a multifold expansion at a new facility—is due by March 2027. Transcript +4
Aeroflex's growth comes from a different angle: keeping those AI chips from overheating. As AI and high-performance computing push rack densities higher, conventional air cooling is reaching its limits. Aeroflex's investor presentation cites a 33.2% CAGR for the global liquid cooling market from CY25 to CY32e, with direct-to-chip cooling growing at 35%. Investor Presentation
The company's liquid cooling skid business has exploded from 46 units in Q3 FY26 to 1,475 units in Q2 FY27, generating ₹50 crore in revenue—about 24% of total sales. Management estimates roughly 40 skids are needed per megawatt of liquid-cooled IT load, linking this to India's roughly 9 GW data centre pipeline. Investor Presentation +2
Aeroflex works exclusively with one large US multinational for liquid cooling in India, supplying through them rather than bidding directly to data centre operators. This exclusive arrangement has been a growth engine but also creates concentration risk. The company has about three months of confirmed supply visibility and doesn't disclose order book values due to customer confidentiality. Transcript +1
To capture this opportunity, Aeroflex is investing ₹202 crore in capacity expansion, targeted for completion by March 2028. This includes ₹124 crore for hoses and braiding (expanding capacity from 20 to 30 million metres annually) and ₹78 crore for liquid cooling skid assemblies (taking capacity from 12,000 to 30,000 units). Management expects this to generate ₹500–550 crore of incremental revenue at peak utilisation. Other +3
Both companies are showing impressive operational improvements alongside revenue growth.
MTAR's EBITDA margin nearly doubled from about 12.5% in Q2 FY26 to 24.4% in Q2 FY27. Interestingly, this wasn't driven by gross margin—which actually fell from 51.2% to 46.8%—but by operating leverage. Employee costs dropped to about 12% of revenue from 25% a year earlier, while other expenses fell to 10% from 14%. Essentially, revenue tripled on a largely fixed cost base. Investor Presentation
Aeroflex achieved a 93 basis point margin expansion to 24.27% in Q2 FY27, attributing this to higher contribution from value-added products and improved operating leverage. Assemblies (which earn 22–26% margins) now make up about 63% of the business, compared to hose-only products at 16–20% margins. Investor Presentation +2
Working capital management has been a game-changer for MTAR. Net working capital days plummeted from 172 for full FY26 to 59 in Q1 FY27 and roughly 50 by September 2026. Receivable days fell from 140 to 68, though inventory days rose to 162. This working capital release, combined with better customer payment terms, supported strong operating cash flow—₹247.7 crore in Q1 FY27 alone. Transcript +3
The biggest risk for both companies is customer concentration.
MTAR doesn't disclose the exact share of its order book from its fuel cell customer, but Clean Energy's 66.7% segment share gives a sense of the exposure. The company is a qualified supplier with deep process capabilities—Nadcap accreditation for over 30 special processes—but pricing negotiations have pressured margins in the past. A Taiwanese competitor also supplies hot boxes to the same customer, limiting MTAR's pricing power. Investor Presentation +3
Aeroflex's liquid cooling business is similarly concentrated. One US multinational accounts for roughly 25–26% of total sales (as of Q4 FY26), and the company dedicates exclusive capacity to them. While this has driven explosive growth, it also means Aeroflex's revenue is tied to one customer's ordering decisions and qualification timelines. Transcript +1
Both companies are working to diversify. MTAR has received ₹45 crore of first-article orders from SLB for sheet-metal components and assemblies for data centre infrastructure, with management citing a potential pipeline about 8x this amount. Aeroflex is expanding internationally, with talks progressing for skid supply beyond India and hose assemblies already serving European data centres. Investor Presentation +3
The return profiles tell an interesting story. MTAR's ROCE declined from about 24% in FY23 to roughly 17% in FY26 (presentation figures) as margins compressed and capital deployed ahead of earnings. Aeroflex's ROCE fell from 37% in FY23 to 20.4% in FY26, pressured by IPO proceeds expanding the equity base and working capital build-up. Investor Presentation +3
Yet both stocks trade at premium multiples. MTAR's current P/E is around 150x trailing earnings, while Aeroflex trades at about 76x [stock_agent]. These valuations assume the growth stories continue and that capex delivers promised returns.
Management guidance is bullish. MTAR targets 80% revenue growth in FY27 with a 24% EBITDA margin, aiming for at least 4–5x asset turnover on new investments. Aeroflex guides to about ₹715–725 crore revenue in FY27 (roughly 60% growth) and sees margins around 25% at peak utilisation. Investor Presentation +3
The critical question is whether these companies can transition from short-cycle, customer-dependent growth to sustainable, diversified businesses. The AI data centre boom provides a powerful tailwind, but the concentration risk is real. As one analyst put it on Aeroflex's call, the company is essentially betting that the air-to-liquid cooling shift is structural, not cyclical. Investor Presentation
For investors, the trade-off is clear: explosive growth and significant upside potential, balanced against customer concentration and execution risk. The next few quarters—particularly as Phase 2 and 3 capacity comes online at MTAR and Aeroflex's expansion progresses—will be crucial in determining whether these precision engineering specialists can sustain their momentum or become cautionary tales about betting too heavily on one customer's capex cycle.